The Inception Team at NSURL-2019 Task 8: Semantic Question Similarity in Arabic

TL;DR

This paper presents an ensemble approach using multilingual BERT for Arabic semantic question similarity, achieving top performance in the NSURL-2019 task with F1-scores up to 96%.

Contribution

It introduces an effective ensemble model leveraging multilingual BERT for Arabic question similarity, outperforming previous methods in the NSURL-2019 challenge.

Findings

Achieved up to 96% F1-score on the dataset.

Ensemble of multilingual BERT models ranked first.

Demonstrated effectiveness of pre-trained multilingual models for Arabic NLP.

Abstract

This paper describes our method for the task of Semantic Question Similarity in Arabic in the workshop on NLP Solutions for Under-Resourced Languages (NSURL). The aim is to build a model that is able to detect similar semantic questions in the Arabic language for the provided dataset. Different methods of determining questions similarity are explored in this work. The proposed models achieved high F1-scores, which range from (88% to 96%). Our official best result is produced from the ensemble model of using a pre-trained multilingual BERT model with different random seeds with 95.924% F1-Score, which ranks the first among nine participants teams.